Gestational diabetes mellitus (GDM) is associated with increased risk of several adverse infant and maternal outcomes and its clinical recognition can reduce these risks. There is concern that the current criteria for GDM may be to restrictive and that residual excess risk of perinatal complications exists below present cutoff values. The proposed study will evaluate whether among women without GDM (as defined by current criteria), increasing levels of maternal glycemia are associated with increased risk of selected perinatal complications: infant severe macrosomia, severe hyperbilirubinemia, hypoglycemia, respiratory distress syndrome, and maternal preeclampsia/eclampsia. To accomplish this, the investigators propose to conduct five nested case-control studies, one for each of the complications of interest, within a large multiethnic cohort of approximately 74,000 women who were screened at 24 to 28 weeks of gestation at Kaiser Permanente, Northern California between 1995 and 1998. In this setting nearly 94 percent of the pregnant women are screened for GDM by a 50 gm., 1 hr. oral glucose tolerance test (50 g, 1-h OGTT) and approximately 15 percent have are abnormal screening and go on to receive the diagnostic (100-g, 3-h OGTT) test. Potential cases of each type of complication will be identified by searching computerized hospitalization and laboratory databases. For each of the infant complications, 500 cases will be randomly selected without knowledge of the maternal glucose values. A single control group for the infant complication case groups will be 1,000 infants randomly selected from all births and frequency matched on gestational age to the distribution of the combined case group. Five hundred women with either preeclampsia or eclampsia and 500 age-matched controls will be randomly selected. The medical records of the 3,000 mother-infant pairs in the four case-control studies on infant complications, and 1,000 women for the case-control study of preeclampsia/eclampsia, will be abstracted to confirm eligibility, and, if so, to ascertain data on possible maternal and infant covariates. Logistic regression will be used to estimate the odds ratios associated with several levels of pregnancy glycemia and perinatal complications. The investigators state that the proposed study will provide important knowledge about the magnitude of the risk of severe perinatal complications associated with degrees of maternal hyperglycemia below the glucose cutpoints currently used to diagnose GDM.